Mining ship deficiency correlations from historical port state control (PSC) inspection data.

Area of Science:

• Maritime Safety
• Data Mining
• Logistics Management

Background:

• Ship deficiencies pose risks to maritime safety, traffic efficiency, and fuel consumption.
• Existing research primarily uses statistical models to link physical deficiencies to Port State Control (PSC) inspection outcomes.
• A gap exists in understanding correlations between parent ship deficiency categories and their subcategories.

Purpose of the Study:

• To explore intrinsic mutual correlations among ship deficiencies using an improved Apriori model.
• To analyze these correlations considering ship property indicators like type, age, deadweight, and gross tonnage.
• To provide guidelines for enhancing PSC inspection efficiency and maritime safety.

Main Methods:

• An improved Apriori algorithm was employed to discover association rules within ship deficiency data.
• Ship property indicators (ship type, age, deadweight, gross tonnage) were integrated into the analysis.
• The PSC inspection dataset was utilized to mine correlations between deficiency categories and subcategories.

Main Results:

• The improved Apriori model successfully identified significant correlations among various ship parent deficiency categories and their subcategories.
• Ship property indicators were found to influence the prevalence and type of deficiencies observed.
• Specific patterns of co-occurring deficiencies were uncovered, offering insights into systemic issues.

Conclusions:

• The study provides a novel approach to uncovering complex relationships within ship deficiency data.
• Findings offer actionable insights for Port State Control (PSC) authorities to optimize inspection strategies.
• Enhanced understanding of deficiency correlations can lead to improved maritime safety and operational efficiency.